Isn't it strange that you can make better PCBs at home than factories make?

[MaximRecoil]

Yes I agree with free_electron until the vector part and with helius for the vector part.

Out of all of the prior errors he made, and which you admittedly agree with him on, this one was my favorite:

To make the screen for printing you need a transparency , expose the screen to uv light via the drawing to cure the resin. then you need to develop the screen in a chemical and wash it. then you put int in the printing press. then you apply ink , then you need to align the screen to the board , then you squeegee the ink through the screen , take the screen off , let the ink dry ( or in case of photo ink : cure it ).

Since you agree with him, you should be able to elaborate on the "cure the resin" and "then you need to develop the screen in a chemical" parts. Both of those things are of course, pure nonsense, but since you believe that you know better, I'd like to hear what you have to say about it. Please include things like a link to where you can buy this "resin" that's intended to be coated onto a screen and cured, and a link to this mystery chemical that you use to "develop the screen", as well as an explanation of what "develop the screen" even means.

I'd also like to know how to reclaim a screen that's been stuffed full of resin which is then cured, given that coating a screen with resin and curing it will turn your screen into a rugged fiber-reinforced composite material that's impervious to most any chemical you could throw at it. Any chemical that could dissolve cured resin probably wouldn't be too friendly to the polyester mesh itself.

[ehughes]

Good luck with those rivets!

I think I'll let someone else with controlled processes drill and plate a .012" hole as well as apply an LPI mask and legend.    I am pretty sure it will always yeild a better result than investing in equipment for home.

[slicendice]

Yes, preparing a PCB for etching and preparing a silk screen (in it's original meaning/terms) is essentially the same process. Chemicals may vary but essentially all the steps are the same. But why do we keep fighting over such things? Who cares? (well those who just want to be right and ignores the original topic of this thread of course :-) )

And those darn Bezier curves...will the electrons fall off the PCB if I don't use complex equations to smooth the traces indefinitely? Com'on guys.

[rx8pilot]

As a (nearly former)  machinist, we use lines and arc to make incredibly complex 3D geometry for 5 axis milling. It is not a limitation even for cosmetically sensitive designs.

A PCBs primary function is to mount components and move the signals around. Artistic curves don't make a better PCB.

Rivets can loosen over time. The FR4 will likely expand and contract much more than the brass. It may not fall out, but provide a less than perfect signal transfer. Either way, I have researched just about every possible way to make PCBs at home and every time I cannot see it being anything more than an interesting experience. The majority of what I design is 4 or more layers which is an immediate disqualification for home brew PCB.

Sent from my horrible mobile....

[slicendice]

As a (nearly former)  machinist, we use lines and arc to make incredibly complex 3D geometry for 5 axis milling. It is not a limitation even for cosmetically sensitive designs.

A PCBs primary function is to mount components and move the signals around. Artistic curves don't make a better PCB.

Rivets can loosen over time. The FR4 will likely expand and contract much more than the brass. It may not fall out, but provide a less than perfect signal transfer. Either way, I have researched just about every possible way to make PCBs at home and every time I cannot see it being anything more than an interesting experience. The majority of what I design is 4 or more layers which is an immediate disqualification for home brew PCB.

Sent from my horrible mobile....

Now we're talking! Nice!

[bpiphany]

Yes, "vector bezier curves", as in, a mathematical representation of a bezier curve which is contained in a vector file, as opposed to a pixels-on-a-grid representation of a bezier curve (raster). When you export an .AI file to .DXF, curves get converted into a series of straight lines rather than remaining as a mathematical representation of curves. I've already posted a screen shot of what that looks like, and here it is again:



It looks like crap because there aren't enough straight lines used for a given distance, to represent the curves. If DXF and Gerber supported vector bezier curves (see above to find out what that bit of common-speak means), then it could remain as a vector bezier curve until the output stage, at which point it would be rasterized, and the rasterization can be at any DPI that the output device is capable of, and with a high enough DPI, they will appear to be smooth curves to the naked eye.

Also, your tacit concession on the matter of your laser-printed foil idea, which you conveniently "forgot" to address, is noted.

Why are you using such a crappy tool that isn't capable of approximating your curves with enough line segments, or even arcs for that matter.


Also see: https://www.boldport.com/blog

[MaximRecoil]

Why are you using such a crappy tool that isn't capable of approximating your curves with enough line segments, or even arcs for that matter.

Your question is meaningless without pointing out a tool that can convert an .AI file to .DXF, and is "capable of approximating [my] curves with enough line segments, or even arcs for that matter."

As a (nearly former)  machinist, we use lines and arc to make incredibly complex 3D geometry for 5 axis milling. It is not a limitation even for cosmetically sensitive designs.

Can you draw the Coca-Cola script logo using only lines and arcs?

Rivets can loosen over time. The FR4 will likely expand and contract much more than the brass. It may not fall out, but provide a less than perfect signal transfer.

The following post seems to be at odds with your theory:

Just my few pfennigs about rivets, is that pretty every CRT TV set made in the last 2 decades ( especially those from Samsung and LG, who were pretty much the biggest volume manufacturers, as they made almost all of the OEM branded ones as well) have had riveted LOPT connections, along with rivets on certain connectors, like those for the scan coils and some of the larger monitors used them in the power supply and on the flyback tuning capacitors as well. All places with high current flows, and with large items soldered to the board, and where you had notorious reliability problems from cracking joints arcing into failure of the line driver transistor ( and many of those transistors, along with the efficiency diode and the base shorting resistor as well) with a bang and in some cases a big burned area of board there.

Those connections were a thing I almost always reflowed on any monitor I had open, for any reason, as they were often either dry jointed, or cracked, on those that were not riveted.

The riveted ones still cracked, but those failures were a lot less common, only occurring on monitors I knew had been dropped, which physically cracked the board as well.

[Codemonkey]

Can you draw the Coca-Cola script logo using only lines and arcs?

What is your obsession with the coca cola logo ?

Any half decent PCB software can just import a bitmap containing any logo you want, including the coca cola one. And yes, you can draw the script using lines and arcs, just lots and lots of very small ones. It's really no big deal.

[slicendice]

Seriously...do I have to code a simple program that can approximate any shape using lines and arcs only, before this nonsense ends?  Only limiting factor would be the physical properties of an atom when applying that data to a real object. Is that accurate enough?

Actually, never mind, it's a complete waste of time.

@MaximRecoil, what is your point of this thread? To pick a fight or to have a decent conversion about what makes up a good PCB, including reasons why the home made PCB  is/can be better than a manufacturer makes?

An infinitely accurate Coca Cola logo on a PCB does not make that particular PCB any better, that's for sure, and anyone could do it.

[slicendice]

@Codemonkey

Exactly!!!

[Arjan Emm]

I do have to say you really carefully choose your forum name. Any reply to your comments produces maximum recoil from you. Recoil is usually an undesirable reaction from a desirable action.

[MaximRecoil]

Can you draw the Coca-Cola script logo using only lines and arcs?

What is your obsession with the coca cola logo ?

Is that a joke, or can you truly not figure that out for yourself? It is an example of a design containing bezier curves that everyone is familiar with.

Any half decent PCB software can just import a bitmap containing any logo you want, including the coca cola one.

This kludge of a workaround has already been discussed.

And yes, you can draw the script using lines and arcs, just lots and lots of very small ones. It's really no big deal.

Not a big deal? How about you give that idea a try and post the results?

Seriously...do I have to code a simple program that can approximate any shape using lines and arcs only, before this nonsense ends?  Only limiting factor would be the physical properties of an atom when applying that data to a real object. Is that accurate enough?

Do it. You might find it isn't as simple as you suppose. Adobe Illustrator can't even export an .AI to .DXF that's particularly compatible with other programs. AutoCAD can usually open and display them correctly, but often runs into issues when extruding to 3D. DipTrace doesn't like them at all; they just show up as random garbage. But maybe you can do better.

Actually, never mind, it's a complete waste of time.

It's a waste of time to write a better AI-to-DXF program than what's already out there? pdf2cad is the best one I've ever used; it at least can convert to polylines and its resulting DXF files don't seem to have any compatibility issues in other programs. However, it still gives a crude straight-line representation of curves. But you could do better, if it weren't a "complete waste of time", right?

@MaximRecoil, what is your point of this thread? To pick a fight or to have a decent conversion about what makes up a good PCB, including reasons why the home made PCB  is/can be better than a manufacturer makes?

Ah, you're a revisionist. I haven't picked any "fights" (arguments). The person who "picks a fight" (starts an argument) is the person who first disputes something that has been said. That came in reply #1, and again in reply #2, and again in reply #3, and so on.

An infinitely accurate Coca Cola logo on a PCB does not make that particular PCB any better

Unlimited drawing functionality is better than limited drawing functionality.

and anyone could do it.

Do it. By the way, I'd be willing to wager a few bucks that you couldn't manually draw an acceptably accurate Coca-Cola logo even in a program which has the right tools for the job, such as Adobe Illustrator. I'd feel confident making that bet because I know there's no easy way for you to cheat, i.e., none of the Coca-Cola logo vector files available for download on the internet are any good when examined closely. The same applies to most free vector files out there.

I do have to say you really carefully choose your forum name. Any reply to your comments produces maximum recoil from you. Recoil is usually an undesirable reaction from a desirable action.

"Maximum"? Where are you seeing that? Is there something wrong with your monitor? The name "Maxim" has no relation to the adjective "maximum", aside from a coincidentally shared sequence of five letters. My username refers to a guy who was from my hometown, by the name of Hiram Maxim, who invented the machine gun, which was recoil-operated.

[Arjan Emm]

Quote from: Arjan Emm on Today at 01:48:45 AM

    I do have to say you really carefully choose your forum name. Any reply to your comments produces maximum recoil from you. Recoil is usually an undesirable reaction from a desirable action.


"Maximum"? Where are you seeing that? Is there something wrong with your monitor? The name "Maxim" has no relation to the adjective "maximum", aside from a coincidentally shared sequence of five letters. My username refers to a guy who was from my hometown, by the name of Hiram Maxim, who invented the machine gun, which was recoil-operated.
« Last Edit: Today at 04:23:57 AM by MaximRecoil »

I rest my case.

[Codemonkey]

I've been designing PCB's on & off professionally for about 27 years now.

I've designed countless PCB's that have been put into production.

I've never once needed to draw a perfect bezier curve.

Go ahead and spend ages pissing about trying to produce your perfect bezier curves on your pretty board with its photorealistic silksceen and silly rivets. Now, get it produced in 10K+ volumes reliably for less than the cost of a perfectly functional board from even the shitest chinese sweatshop. Good luck with that.

Then stick your mega expensive pretty PCB in a box where it can't be seen anyway!

Your argument is utterly pointless.

What a numpty 

[MaximRecoil]

I rest my case.

Your resignation is accepted.

I've been designing PCB's on & off professionally for about 27 years now.

I've designed countless PCB's that have been put into production.

I've never once needed to draw a perfect bezier curve.

Thanks for the anecdote. It is irrelevant though.

Go ahead and spend ages pissing about trying to produce your perfect bezier curves on your pretty board with its photorealistic silksceen and silly rivets.

^^^ Yet another person who has replied to a thread that he apparently didn't read. As I've said several times, I have no intention of making a PCB of any kind, for any reason, at home. In any case, it wouldn't take "ages" for me to do it, nor would there be any "pissing about", because I know exactly how to produce perfect bezier curves, a photorealistic screen print, and install rivets.

silly rivets

Because stronger is "silly", right? Once again:

Just my few pfennigs about rivets, is that pretty every CRT TV set made in the last 2 decades ( especially those from Samsung and LG, who were pretty much the biggest volume manufacturers, as they made almost all of the OEM branded ones as well) have had riveted LOPT connections, along with rivets on certain connectors, like those for the scan coils and some of the larger monitors used them in the power supply and on the flyback tuning capacitors as well. All places with high current flows, and with large items soldered to the board, and where you had notorious reliability problems from cracking joints arcing into failure of the line driver transistor ( and many of those transistors, along with the efficiency diode and the base shorting resistor as well) with a bang and in some cases a big burned area of board there.

Those connections were a thing I almost always reflowed on any monitor I had open, for any reason, as they were often either dry jointed, or cracked, on those that were not riveted.

The riveted ones still cracked, but those failures were a lot less common, only occurring on monitors I knew had been dropped, which physically cracked the board as well.

Samsung and LG; they're just so "silly".

Now, get it produced in 10K+ volumes reliably for less than the cost of a perfectly functional board from even the shitest chinese sweatshop. Good luck with that.

What do "10K+ volumes" have to do with making PCBs at home? Nothing, that's what.

Then stick your mega expensive pretty PCB

There's nothing expensive about making a PCB the way I described if you already have the tools you need, let alone "mega" expensive. The rivets are perhaps the most expensive part (depending on how many of them you need to use), and those are about 3 cents each from here, and perhaps cheaper elsewhere.

in a box where it can't be seen anyway!

How do you know where the PCB will be mounted? A lot of people who do electronics projects at home use clear enclosures so that the internals can be seen.

Your argument is utterly pointless.

Given all of your false premises above, your conclusion based on them is inherently invalid, and can legitimately be dismissed out of hand.

What a numpty 

So says the fellow who just made a slew of false, irrelevant, and otherwise invalid statements.

[Richard Crowley]

Nominated for Stupidest Thread of the Year for 2016.   

[Codemonkey]

I have no intention of making a PCB of any kind, for any reason, at home. In any case, it wouldn't take "ages" for me to do it, nor would there be any "pissing about", because I know exactly how to produce perfect bezier curves, a photorealistic screen print, and install rivets.

One again, bezier curves - totally not needed on a PCB.
Photorealistic screen print - totally not needed by anyone ever on a PCB.
Rivets - installed tediously one by one rather than electroplate the whole lot at once.

What do "10K+ volumes" have to do with making PCBs at home? Nothing, that's what.

You're the one claiming to be able to produce better PCB's at home than a factory. Factory's produce billions of PCB's per year reliably incredibly cheaply. Have you ever produced a single PCB ? If so, go on, lets all see your better than a factory can do PCB ? (despite claiming above you never even intend ever producing even a single one at home).

There's nothing expensive about making a PCB the way I described if you already have the tools you need, let alone "mega" expensive. The rivets are perhaps the most expensive part (depending on how many of them you need to use), and those are about 3 cents each from here, and perhaps cheaper elsewhere.

Attached is a photo of a simple PCB I knocked up for a home project. On it there are 182 plated through holes and only one through hole component. By far, the majority of holes are used stitch ground planes together to improve the EMC performance of the PCB itself since the board has an on board 2.4GHz radio. Cost of using rivets using your figures: ~$5.46, not including labour time to fit them, per PCB. I got 10 boards made in China, shipped half way round the world to the UK for about $20. I've populated 3 so far, all work perfectly.

The NXP RF module in the middle (which is only about the size of the end of my thumb), has well over 100 plated through holes, just in the bit of board you can see that isn't covered by the screening can, all in an area less than the size of my thumb nail. Again, no through hole components, so strength isn't an issue, but they didn't just put them all there for a bit of a laugh.

They all serve a purpose to form a good RF conduit from the IC under the can to the antenna on the end (that track with loads of vias either side is I believe what is known as coplanar waveguide), and to stop bits of the ground plane acting as antenna's and radiating any unwanted harmonics.

I can buy those modules from Farnell in the UK for £6.63 each (~$8.19) in one offs. If using your method, the cost of the rivets alone would be $3, plus then there is the cost of the rest of the bare PCB (looks to be 4 layers) including loads of time (time = money) to fit the many many individual rivets, the IC, passives, crystal and the screening can that they all sit under, plus placing them, soldering them, testing the finished module, packaging them in tape & reel format, and shipping them half way round the world to the UK (according to the page on the Farnell website, they're made in the far east). I doubt very much that your "better" method would allow them to be so cheap whilst still allowing both NXP and Farnell to be able to make a profit on them.

How would your proposed "better than a factory can make" method cope with this ? You wouldn't even be able to produce a single board at home that could perform as well as the one that the module uses.

So says the fellow who just made a slew of false, irrelevant, and otherwise invalid statements.

Says the muppet who defines a better PCB to be one containing utterly pointless bezier curves and photorealistic silk screens.

Have you ever designed and produced even a single PCB ? You're the one that started the thread by making dumbest claim of the year!

I've yet to see any one of the many industry professionals, or even hobbyists that have posted in this thread agree with you.

Do jog on... 

[slicendice]

@Codemonkey

That's one really good looking PCB! Not the best of soldering results for the tricky and tiny surface mount components, but if it works, then great.

I could not do it any better, probably would look a lot worse than yours. 

Great job!

[Codemonkey]

Lol, thanks. I used a cheap paste stencil from https://www.smtstencil.co.uk, hand placed the components and reflowed the board in a toaster oven with a DIY temperature controller.

The dodgy soldering is just where i've tacked on test points and changed things (this was the first board to get populated so its had scopes and logic analyser hanging off it while I did software development).

How it ever worked without any magic bezier curves though is a complete mystery 

[slicendice]

How it ever worked without any magic bezier curves though is a complete mystery 

Yeah, amazing, one would think the electrons would fall off the sharp edges. You must have magic electrons in your coin cell battery. 

[MaximRecoil]

One again, bezier curves - totally not needed on a PCB.
Photorealistic screen print - totally not needed by anyone ever on a PCB.
Rivets - installed tediously one by one rather than electroplate the whole lot at once.

I never said anything about "need" (as I've pointed out multiple times already), and tedium is irrelevant. As such, this is a non sequitur from you; consider it dismissed as such.

You're the one claiming to be able to produce better PCB's at home than a factory. Factory's produce billions of PCB's per year reliably incredibly cheaply. Have you ever produced a single PCB ? If so, go on, lets all see your better than a factory can do PCB ? (despite claiming above you never even intend ever producing even a single one at home).

This red herring from you = your tacit agreement that "10K+ volumes" has nothing to do with making PCBs at home.

Attached is a photo of a simple PCB I knocked up for a home project. On it there are 182 plated through holes and only one through hole component. By far, the majority of holes are used stitch ground planes together to improve the EMC performance of the PCB itself since the board has an on board 2.4GHz radio. Cost of using rivets using your figures: ~$5.46, not including labour time to fit them, per PCB. I got 10 boards made in China, shipped half way round the world to the UK for about $20. I've populated 3 so far, all work perfectly.

$5.46 added to the cost is not expensive, much less "mega expensive". If I were to make a PCB at home, I'd only be making one example of any given design.

The NXP RF module in the middle (which is only about the size of the end of my thumb), has well over 100 plated through holes, just in the bit of board you can see that isn't covered by the screening can, all in an area less than the size of my thumb nail. Again, no through hole components, so strength isn't an issue, but they didn't just put them all there for a bit of a laugh.

They all serve a purpose to form a good RF conduit from the IC under the can to the antenna on the end (that track with loads of vias either side is I believe what is known as coplanar waveguide), and to stop bits of the ground plane acting as antenna's and radiating any unwanted harmonics.

I can buy those modules from Farnell in the UK for £6.63 each (~$8.19) in one offs. If using your method, the cost of the rivets alone would be $3, plus then there is the cost of the rest of the bare PCB (looks to be 4 layers) including loads of time (time = money) to fit the many many individual rivets, the IC, passives, crystal and the screening can that they all sit under, plus placing them, soldering them, testing the finished module, packaging them in tape & reel format, and shipping them half way round the world to the UK (according to the page on the Farnell website, they're made in the far east). I doubt very much that your "better" method would allow them to be so cheap whilst still allowing both NXP and Farnell to be able to make a profit on them.

How would your proposed "better than a factory can make" method cope with this ? You wouldn't even be able to produce a single board at home that could perform as well as the one that the module uses.

I never said "better than a factory can make", and this isn't the first time I've pointed out that I never said that. Second, I've said more than once "within certain limitations". It doesn't matter how many examples you give of specific PCB designs that aren't particularly suited to making at home, because there are literally an unlimited number of designs which are suited to making at home, both designs which have already been conceived and designs which are yet to be conceived.

Says the muppet

Comical Irony Alert, and Monkey See, Monkey Do.

who defines a better PCB to be one containing utterly pointless bezier curves and photorealistic silk screens.

Given that you conveniently "forgot" to include stronger through-holes / vias, this is a straw man from you. And yes, less limitations for a given design is better than more limitations.

Have you ever designed

Yes, one, but the other PCBs I've drawn up weren't my design, they were reproductions of old PCBs, and I've run into an annoying limitation of PCB design software, i.e., no way to natively draw a bezier curve to replicate the curvy traces that were common with old hand-drawn PCBs.

and produced even a single PCB ?

No, and I have no intentions of doing so. This is yet another thing I've already said multiple times, and at this point it is safe to say that you don't read so well.

You're the one that started the thread by making dumbest claim of the year!

You just tacitly (and unwittingly) claimed that you and every other naysayer who has posted on this thread is an idiot, given that none of you have been able to refute the alleged "dumbest claim of the year".

I've yet to see any one of the many industry professionals, or even hobbyists that have posted in this thread agree with you.

Facts aren't up for a vote. In fact, they aren't even debatable, but don't let that stop you from trying. The fact that nearly everything you have posted is a non sequitur, with a good portion of those non sequiturs being attributable to a reading comprehension failure, makes you particularly easy to argue with.

[G0HZU]

A good PCB manufacturer will manufacture a PCB to YOUR requirements/specs.

If you need thick copper weight (eg 4 oz) to improve current handling and connector reliability they can often do this for you. Presumably you can order PCBs with rivets if you like?

Or maybe you think those Samsung and LG TVs used 'homemade' PCBs to get them to feature strengthening rivets in their TVs?

A suitable PCB factory will make the board to your requirements. If you deliberately ask them to make a board that doesn't meet the requirements for strength/thermal/current rating or you send them crude drawings rather than fine resolution drawings for the artworks then who is at fault here?

You might as well argue that if you send the PCB factory an old issue of artwork (that doesn't have all the required reliability features on it or the latest resolution) then you can do a better job at home if you make it to the latest issue.

[G0HZU]

The Bezier curve argument is just as silly. You might as well argue that the PCB factory can't replicate an 'arty' PCB that shows your palm print dipped in liquid and slapped on a PCB complete with random splashes and various other arty anomalies. They can't replicate your etching anomalies either and reproduce the true 'art' of the homemade original. But who cares?

[mrpackethead]

Facts aren't up for a vote.

Then how did Donaldn Trump get elected?   I'm going to assume Mr Recoil ( its not a fact, just an assumption ) that you are resonably intellgent. Perhaps a very High IQ.    However, your showing an exceptionally low EQ, which is the ability to deal with people.   Your ability to make a difference for good in this world is a combination of both.

Your net worth to the world could be massively increased, if you just took a little time to figure out how people are perceiving you.

[slicendice]

Happy new year to everybody in this thread too...!